UnityShip: A Large-Scale Synthetic Dataset for Ship Recognition in Aerial Images

As a data-driven approach, deep learning requires a large amount of annotated data for training to obtain a sufficiently accurate and generalized model, especially in the field of computer vision. However, when compared with generic object recognition datasets, aerial image datasets are more challenging to acquire and more expensive to label. Obtaining a large amount of high-quality aerial image data for object recognition and image understanding is an urgent problem. Existing studies show that synthetic data can effectively reduce the amount of training data required. Therefore, in this paper, we propose the first synthetic aerial image dataset for ship recognition, called UnityShip. This dataset contains over 100,000 synthetic images and 194,054 ship instances, including 79 different ship models in ten categories and six different large virtual scenes with different time periods, weather environments, and altitudes. The annotations include environmental information, instance-level horizontal bounding boxes, oriented bounding boxes, and the type and ID of each ship. This provides the basis for object detection, oriented object detection, fine-grained recognition, and scene recognition. To investigate the applications of UnityShip, the synthetic data were validated for model pre-training and data augmentation using three different object detection algorithms and six existing real-world ship detection datasets. Our experimental results show that for small-sized and medium-sized real-world datasets, the synthetic data achieve an improvement in model pre-training and data augmentation, showing the value and potential of synthetic data in aerial image recognition and understanding tasks.

A multi-source information fusion method for ship target recognition based on bayesian inference and evidence theory

The ship target recognition (STR) is greatly related to the battlefield situation awareness, which has recently gained prominence in the military domains. With the diversification and complexity of military missions, ship targets are mostly performed in the form of formations. Therefore, using the formation information to improve the accuracy of the ship target type recognition is worth studying. To effectively identify ship target type, we in this paper jointly consider the ship dynamic, formation, and feature information to propose a STR method based on Bayesian inference and evidence theory. Specifically, we first calculate the ship position distance matrix and the directional distance matrix with the Dynamic Time Warping (DTW) and the difference-vector algorithm taken into account. Then, we use the two distance matrices to obtain the ship formation information at different distance thresholds by the hierarchical clustering method, based on which we can infer the ship type. Thirdly, formation information and other attribute information are as nodes of the Bayesian Network (BN) to infer the ship type. Afterward, we can convert the recognition results at different thresholds into body of evidences (BOEs) as multiple information sources. Finally, we fuse the BOEs to get the final recognition. The proposed method is verified in simulation battle scenario in this paper. The simulation results demonstrate that the proposed method achieves performance superiority as compared with other ship recognition methods in terms of recognition accuracy.

A Hybrid Method for Inland Ship Recognition Using Marine Radar and Closed-Circuit Television

Vessel recognition plays important role in ensuring navigation safety. However, existing methods are mainly based on a single sensor, such as automatic identification system (AIS), marine radar, closed-circuit television (CCTV), etc. To this end, this paper proposes a coarse-to-fine recognition method by fusing CCTV and marine radar, called multi-scale matching vessel recognition (MSM-VR). This method first proposes a novel calibration method that does not use any additional calibration target. The calibration is transformed to solve an N point registration model. Furthermore, marine radar image is used for coarse detection. A region of interest (ROI) area is computed for coarse detection results. Lastly, we design a novel convolutional neural network (CNN) called VesNet and transform the recognition into feature extraction. The VesNet is used to extract the vessel features. As a result, the MVM-VR method has been validated by using actual datasets collected along different waterways such as Nanjing waterway and Wuhan waterway, China, covering different times and weather conditions. Experimental results show that the MSM-VR method can adapt to different times, different weather conditions, and different waterways with good detection stability. The recognition accuracy is no less than 96%. Compared to other methods, the proposed method has high accuracy and great robustness.

Radar High-Resolution Range Profile Ship Recognition Using Two-Channel Convolutional Neural Networks Concatenated with Bidirectional Long Short-Term Memory

Radar automatic target recognition is a critical research topic in radar signal processing. Radar high-resolution range profiles (HRRPs) describe the radar characteristics of a target, that is, the characteristics of the target that is reflected by the microwave emitted by the radar are implicit in it. In conventional radar HRRP target recognition methods, prior knowledge of the radar is necessary for target recognition. The application of deep-learning methods in HRRPs began in recent years, and most of them are convolutional neural network (CNN) and its variants, and recurrent neural network (RNN) and the combination of RNN and CNN are relatively rarely used. The continuous pulses emitted by the radar hit the ship target, and the received HRRPs of the reflected wave seem to provide the geometric characteristics of the ship target structure. When the radar pulses are transmitted to the ship, different positions on the ship have different structures, so each range cell of the echo reflected in the HRRP will be different, and adjacent structures should also have continuous relational characteristics. This inspired the authors to propose a model to concatenate the features extracted by the two-channel CNN with bidirectional long short-term memory (BiLSTM). Various filters are used in two-channel CNN to extract deep features and fed into the following BiLSTM. The BiLSTM model can effectively capture long-distance dependence, because BiLSTM can be trained to retain critical information and achieve two-way timing dependence. Therefore, the two-way spatial relationship between adjacent range cells can be used to obtain excellent recognition performance. The experimental results revealed that the proposed method is robust and effective for ship recognition.